2 * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3 * Copyright (C) $$Year-Rel$$ The Jalview Authors
5 * This file is part of Jalview.
7 * Jalview is free software: you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation, either version 3
10 * of the License, or (at your option) any later version.
12 * Jalview is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty
14 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15 * PURPOSE. See the GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
19 * The Jalview Authors are detailed in the 'AUTHORS' file.
23 import jalview.analysis.AlignmentUtils;
24 import jalview.analysis.SequenceIdMatcher;
25 import jalview.api.AlignViewportI;
26 import jalview.api.FeatureColourI;
27 import jalview.api.FeaturesSourceI;
28 import jalview.datamodel.AlignedCodonFrame;
29 import jalview.datamodel.Alignment;
30 import jalview.datamodel.AlignmentI;
31 import jalview.datamodel.SequenceDummy;
32 import jalview.datamodel.SequenceFeature;
33 import jalview.datamodel.SequenceI;
34 import jalview.io.gff.GffHelperBase;
35 import jalview.io.gff.GffHelperFactory;
36 import jalview.io.gff.GffHelperI;
37 import jalview.schemes.FeatureColour;
38 import jalview.schemes.UserColourScheme;
39 import jalview.util.MapList;
40 import jalview.util.ParseHtmlBodyAndLinks;
41 import jalview.util.StringUtils;
43 import java.io.IOException;
44 import java.util.ArrayList;
45 import java.util.Arrays;
46 import java.util.HashMap;
47 import java.util.Iterator;
48 import java.util.List;
50 import java.util.Map.Entry;
51 import java.util.StringTokenizer;
54 * Parses and writes features files, which may be in Jalview, GFF2 or GFF3
55 * format. These are tab-delimited formats but with differences in the use of
58 * A Jalview feature file may define feature colours and then declare that the
59 * remainder of the file is in GFF format with the line 'GFF'.
61 * GFF3 files may include alignment mappings for features, which Jalview will
62 * attempt to model, and may include sequence data following a ##FASTA line.
69 public class FeaturesFile extends AlignFile implements FeaturesSourceI
71 private static final String ID_NOT_SPECIFIED = "ID_NOT_SPECIFIED";
73 private static final String NOTE = "Note";
75 protected static final String TAB = "\t";
77 protected static final String GFF_VERSION = "##gff-version";
79 private AlignmentI lastmatchedAl = null;
81 private SequenceIdMatcher matcher = null;
83 protected AlignmentI dataset;
85 protected int gffVersion;
88 * Creates a new FeaturesFile object.
95 * Constructor which does not parse the file immediately
101 public FeaturesFile(String inFile, String type) throws IOException
103 super(false, inFile, type);
108 * @throws IOException
110 public FeaturesFile(FileParse source) throws IOException
116 * Constructor that optionally parses the file immediately
118 * @param parseImmediately
121 * @throws IOException
123 public FeaturesFile(boolean parseImmediately, String inFile, String type)
126 super(parseImmediately, inFile, type);
130 * Parse GFF or sequence features file using case-independent matching,
134 * - alignment/dataset containing sequences that are to be annotated
136 * - hashtable to store feature colour definitions
138 * - process html strings into plain text
139 * @return true if features were added
141 public boolean parse(AlignmentI align,
142 Map<String, FeatureColourI> colours,
145 return parse(align, colours, removeHTML, false);
149 * Extends the default addProperties by also adding peptide-to-cDNA mappings
150 * (if any) derived while parsing a GFF file
153 public void addProperties(AlignmentI al)
155 super.addProperties(al);
156 if (dataset != null && dataset.getCodonFrames() != null)
158 AlignmentI ds = (al.getDataset() == null) ? al : al.getDataset();
159 for (AlignedCodonFrame codons : dataset.getCodonFrames())
161 ds.addCodonFrame(codons);
167 * Parse GFF or Jalview format sequence features file
170 * - alignment/dataset containing sequences that are to be annotated
172 * - hashtable to store feature colour definitions
174 * - process html strings into plain text
175 * @param relaxedIdmatching
176 * - when true, ID matches to compound sequence IDs are allowed
177 * @return true if features were added
179 public boolean parse(AlignmentI align,
180 Map<String, FeatureColourI> colours,
181 boolean removeHTML, boolean relaxedIdmatching)
183 Map<String, String> gffProps = new HashMap<String, String>();
185 * keep track of any sequences we try to create from the data
187 List<SequenceI> newseqs = new ArrayList<SequenceI>();
193 String featureGroup = null;
195 while ((line = nextLine()) != null)
197 // skip comments/process pragmas
198 if (line.length() == 0 || line.startsWith("#"))
200 if (line.toLowerCase().startsWith("##"))
202 processGffPragma(line, gffProps, align, newseqs);
207 gffColumns = line.split("\\t"); // tab as regex
208 if (gffColumns.length == 1)
210 if (line.trim().equalsIgnoreCase("GFF"))
213 * Jalview features file with appended GFF
214 * assume GFF2 (though it may declare ##gff-version 3)
221 if (gffColumns.length > 1 && gffColumns.length < 4)
224 * if 2 or 3 tokens, we anticipate either 'startgroup', 'endgroup' or
225 * a feature type colour specification
227 String ft = gffColumns[0];
228 if (ft.equalsIgnoreCase("startgroup"))
230 featureGroup = gffColumns[1];
232 else if (ft.equalsIgnoreCase("endgroup"))
234 // We should check whether this is the current group,
235 // but at present there's no way of showing more than 1 group
240 String colscheme = gffColumns[1];
241 FeatureColourI colour = FeatureColour
242 .parseJalviewFeatureColour(colscheme);
245 colours.put(ft, colour);
252 * if not a comment, GFF pragma, startgroup, endgroup or feature
253 * colour specification, that just leaves a feature details line
254 * in either Jalview or GFF format
258 parseJalviewFeature(line, gffColumns, align, colours, removeHTML,
259 relaxedIdmatching, featureGroup);
263 parseGff(gffColumns, align, relaxedIdmatching, newseqs);
267 } catch (Exception ex)
269 // should report somewhere useful for UI if necessary
270 warningMessage = ((warningMessage == null) ? "" : warningMessage)
271 + "Parsing error at\n" + line;
272 System.out.println("Error parsing feature file: " + ex + "\n" + line);
273 ex.printStackTrace(System.err);
279 * experimental - add any dummy sequences with features to the alignment
280 * - we need them for Ensembl feature extraction - though maybe not otherwise
282 for (SequenceI newseq : newseqs)
284 if (newseq.getSequenceFeatures() != null)
286 align.addSequence(newseq);
293 * Try to parse a Jalview format feature specification and add it as a
294 * sequence feature to any matching sequences in the alignment. Returns true
295 * if successful (a feature was added), or false if not.
300 * @param featureColours
302 * @param relaxedIdmatching
303 * @param featureGroup
305 protected boolean parseJalviewFeature(String line, String[] gffColumns,
306 AlignmentI alignment, Map<String, FeatureColourI> featureColours,
307 boolean removeHTML, boolean relaxedIdMatching, String featureGroup)
310 * tokens: description seqid seqIndex start end type [score]
312 if (gffColumns.length < 6)
314 System.err.println("Ignoring feature line '" + line
315 + "' with too few columns (" + gffColumns.length + ")");
318 String desc = gffColumns[0];
319 String seqId = gffColumns[1];
320 SequenceI seq = findSequence(seqId, alignment, null, relaxedIdMatching);
322 if (!ID_NOT_SPECIFIED.equals(seqId))
324 seq = findSequence(seqId, alignment, null, relaxedIdMatching);
330 String seqIndex = gffColumns[2];
333 int idx = Integer.parseInt(seqIndex);
334 seq = alignment.getSequenceAt(idx);
335 } catch (NumberFormatException ex)
337 System.err.println("Invalid sequence index: " + seqIndex);
343 System.out.println("Sequence not found: " + line);
347 int startPos = Integer.parseInt(gffColumns[3]);
348 int endPos = Integer.parseInt(gffColumns[4]);
350 String ft = gffColumns[5];
352 if (!featureColours.containsKey(ft))
355 * Perhaps an old style groups file with no colours -
356 * synthesize a colour from the feature type
358 UserColourScheme ucs = new UserColourScheme(ft);
359 featureColours.put(ft, new FeatureColour(ucs.findColour('A')));
361 SequenceFeature sf = new SequenceFeature(ft, desc, "", startPos,
362 endPos, featureGroup);
363 if (gffColumns.length > 6)
365 float score = Float.NaN;
368 score = new Float(gffColumns[6]).floatValue();
369 // update colourgradient bounds if allowed to
370 } catch (NumberFormatException ex)
377 parseDescriptionHTML(sf, removeHTML);
379 seq.addSequenceFeature(sf);
382 && (seq = alignment.findName(seq, seqId, false)) != null)
384 seq.addSequenceFeature(new SequenceFeature(sf));
390 * Process a feature type colour specification
393 * the current input line (for error messages only)
395 * the first token on the line
397 * holds tokens on the line
399 * map to which to add derived colour specification
401 protected void parseFeatureColour(String line, String featureType,
402 String[] gffColumns, Map<String, FeatureColourI> colours)
404 FeatureColourI colour = null;
405 String colscheme = gffColumns[1];
406 if (colscheme.indexOf("|") > -1
407 || colscheme.trim().equalsIgnoreCase("label"))
409 colour = parseGraduatedColourScheme(line, colscheme);
413 UserColourScheme ucs = new UserColourScheme(colscheme);
414 colour = new FeatureColour(ucs.findColour('A'));
418 colours.put(featureType, colour);
423 * Parse a Jalview graduated colour descriptor
426 * @param colourDescriptor
429 protected FeatureColourI parseGraduatedColourScheme(String line,
430 String colourDescriptor)
432 // Parse '|' separated graduated colourscheme fields:
433 // [label|][mincolour|maxcolour|[absolute|]minvalue|maxvalue|thresholdtype|thresholdvalue]
434 // can either provide 'label' only, first is optional, next two
435 // colors are required (but may be
436 // left blank), next is optional, nxt two min/max are required.
437 // first is either 'label'
438 // first/second and third are both hexadecimal or word equivalent
440 // next two are values parsed as floats.
441 // fifth is either 'above','below', or 'none'.
442 // sixth is a float value and only required when fifth is either
443 // 'above' or 'below'.
444 StringTokenizer gcol = new StringTokenizer(colourDescriptor, "|", true);
446 float min = Float.MIN_VALUE, max = Float.MAX_VALUE;
447 boolean labelCol = false;
449 String mincol = gcol.nextToken();
453 .println("Expected either 'label' or a colour specification in the line: "
457 String maxcol = null;
458 if (mincol.toLowerCase().indexOf("label") == 0)
461 mincol = (gcol.hasMoreTokens() ? gcol.nextToken() : null); // skip '|'
462 mincol = (gcol.hasMoreTokens() ? gcol.nextToken() : null);
464 String abso = null, minval, maxval;
467 // at least four more tokens
468 if (mincol.equals("|"))
474 gcol.nextToken(); // skip next '|'
476 // continue parsing rest of line
477 maxcol = gcol.nextToken();
478 if (maxcol.equals("|"))
484 gcol.nextToken(); // skip next '|'
486 abso = gcol.nextToken();
487 gcol.nextToken(); // skip next '|'
488 if (abso.toLowerCase().indexOf("abso") != 0)
495 minval = gcol.nextToken();
496 gcol.nextToken(); // skip next '|'
498 maxval = gcol.nextToken();
499 if (gcol.hasMoreTokens())
501 gcol.nextToken(); // skip next '|'
505 if (minval.length() > 0)
507 min = Float.valueOf(minval);
509 } catch (Exception e)
512 .println("Couldn't parse the minimum value for graduated colour for type ("
514 + ") - did you misspell 'auto' for the optional automatic colour switch ?");
519 if (maxval.length() > 0)
521 max = Float.valueOf(maxval);
523 } catch (Exception e)
526 .println("Couldn't parse the maximum value for graduated colour for type ("
527 + colourDescriptor + ")");
533 // add in some dummy min/max colours for the label-only
539 FeatureColourI colour = null;
542 colour = new FeatureColour(
543 new UserColourScheme(mincol).findColour('A'),
544 new UserColourScheme(maxcol).findColour('A'), min, max);
545 } catch (Exception e)
547 System.err.println("Couldn't parse the graduated colour scheme ("
548 + colourDescriptor + ")");
553 colour.setColourByLabel(labelCol);
554 colour.setAutoScaled(abso == null);
555 // add in any additional parameters
556 String ttype = null, tval = null;
557 boolean hasThreshold = false;
558 if (gcol.hasMoreTokens())
560 // threshold type and possibly a threshold value
561 ttype = gcol.nextToken();
562 if (ttype.toLowerCase().startsWith("below"))
564 colour.setBelowThreshold(true);
567 else if (ttype.toLowerCase().startsWith("above"))
569 colour.setAboveThreshold(true);
574 if (!ttype.toLowerCase().startsWith("no"))
576 System.err.println("Ignoring unrecognised threshold type : "
586 tval = gcol.nextToken();
587 colour.setThreshold(new Float(tval).floatValue());
588 } catch (Exception e)
590 System.err.println("Couldn't parse threshold value as a float: ("
595 // parse the thresh-is-min token ?
596 if (gcol.hasMoreTokens())
599 .println("Ignoring additional tokens in parameters in graduated colour specification\n");
600 while (gcol.hasMoreTokens())
602 System.err.println("|" + gcol.nextToken());
604 System.err.println("\n");
611 * clear any temporary handles used to speed up ID matching
613 protected void resetMatcher()
615 lastmatchedAl = null;
620 * Returns a sequence matching the given id, as follows
622 * <li>strict matching is on exact sequence name</li>
623 * <li>relaxed matching allows matching on a token within the sequence name,
625 * <li>first tries to find a match in the alignment sequences</li>
626 * <li>else tries to find a match in the new sequences already generated while
627 * parsing the features file</li>
628 * <li>else creates a new placeholder sequence, adds it to the new sequences
629 * list, and returns it</li>
635 * @param relaxedIdMatching
639 protected SequenceI findSequence(String seqId, AlignmentI align,
640 List<SequenceI> newseqs, boolean relaxedIdMatching)
642 // TODO encapsulate in SequenceIdMatcher, share the matcher
643 // with the GffHelper (removing code duplication)
644 SequenceI match = null;
645 if (relaxedIdMatching)
647 if (lastmatchedAl != align)
649 lastmatchedAl = align;
650 matcher = new SequenceIdMatcher(align.getSequencesArray());
653 matcher.addAll(newseqs);
656 match = matcher.findIdMatch(seqId);
660 match = align.findName(seqId, true);
661 if (match == null && newseqs != null)
663 for (SequenceI m : newseqs)
665 if (seqId.equals(m.getName()))
673 if (match == null && newseqs != null)
675 match = new SequenceDummy(seqId);
676 if (relaxedIdMatching)
678 matcher.addAll(Arrays.asList(new SequenceI[] { match }));
680 // add dummy sequence to the newseqs list
686 public void parseDescriptionHTML(SequenceFeature sf, boolean removeHTML)
688 if (sf.getDescription() == null)
692 ParseHtmlBodyAndLinks parsed = new ParseHtmlBodyAndLinks(
693 sf.getDescription(), removeHTML, newline);
695 sf.description = (removeHTML) ? parsed.getNonHtmlContent()
697 for (String link : parsed.getLinks())
705 * generate a features file for seqs includes non-pos features by default.
708 * source of sequence features
710 * hash of feature types and colours
711 * @return features file contents
713 public String printJalviewFormat(SequenceI[] sequences,
714 Map<String, FeatureColourI> visible)
716 return printJalviewFormat(sequences, visible, true, true);
720 * generate a features file for seqs with colours from visible (if any)
724 * @param featureColours
725 * hash of Colours for each feature type
727 * when true only feature types in 'visible' will be output
729 * indicates if non-positional features should be output (regardless
731 * @return features file contents
733 public String printJalviewFormat(SequenceI[] sequences,
734 Map<String, FeatureColourI> featureColours, boolean visOnly,
737 StringBuilder out = new StringBuilder(256);
738 boolean featuresGen = false;
739 if (visOnly && !nonpos
740 && (featureColours == null || featureColours.size() < 1))
742 // no point continuing.
743 return "No Features Visible";
746 if (featureColours != null && visOnly)
748 // write feature colours only if we're given them and we are generating
750 // TODO: decide if feature links should also be written here ?
751 Iterator<String> en = featureColours.keySet().iterator();
754 String featureType = en.next();
755 FeatureColourI colour = featureColours.get(featureType);
756 out.append(colour.toJalviewFormat(featureType)).append(newline);
760 // Work out which groups are both present and visible
761 List<String> groups = new ArrayList<String>();
763 boolean isnonpos = false;
765 SequenceFeature[] features;
766 for (int i = 0; i < sequences.length; i++)
768 features = sequences[i].getSequenceFeatures();
769 if (features != null)
771 for (int j = 0; j < features.length; j++)
773 isnonpos = features[j].begin == 0 && features[j].end == 0;
774 if ((!nonpos && isnonpos)
775 || (!isnonpos && visOnly && !featureColours
776 .containsKey(features[j].type)))
781 if (features[j].featureGroup != null
782 && !groups.contains(features[j].featureGroup))
784 groups.add(features[j].featureGroup);
793 if (groups.size() > 0 && groupIndex < groups.size())
795 group = groups.get(groupIndex);
797 out.append("STARTGROUP").append(TAB);
806 for (int i = 0; i < sequences.length; i++)
808 features = sequences[i].getSequenceFeatures();
809 if (features != null)
811 for (SequenceFeature sequenceFeature : features)
813 isnonpos = sequenceFeature.begin == 0 && sequenceFeature.end == 0;
814 if ((!nonpos && isnonpos)
815 || (!isnonpos && visOnly && !featureColours
816 .containsKey(sequenceFeature.type)))
818 // skip if feature is nonpos and we ignore them or if we only
819 // output visible and it isn't non-pos and it's not visible
824 && (sequenceFeature.featureGroup == null || !sequenceFeature.featureGroup
830 if (group == null && sequenceFeature.featureGroup != null)
834 // we have features to output
836 if (sequenceFeature.description == null
837 || sequenceFeature.description.equals(""))
839 out.append(sequenceFeature.type).append(TAB);
843 if (sequenceFeature.links != null
844 && sequenceFeature.getDescription().indexOf("<html>") == -1)
846 out.append("<html>");
849 out.append(sequenceFeature.description);
850 if (sequenceFeature.links != null)
852 for (int l = 0; l < sequenceFeature.links.size(); l++)
854 String label = sequenceFeature.links.elementAt(l);
855 String href = label.substring(label.indexOf("|") + 1);
856 label = label.substring(0, label.indexOf("|"));
858 if (sequenceFeature.description.indexOf(href) == -1)
860 out.append(" <a href=\"" + href + "\">" + label
865 if (sequenceFeature.getDescription().indexOf("</html>") == -1)
867 out.append("</html>");
873 out.append(sequences[i].getName());
874 out.append("\t-1\t");
875 out.append(sequenceFeature.begin);
877 out.append(sequenceFeature.end);
879 out.append(sequenceFeature.type);
880 if (!Float.isNaN(sequenceFeature.score))
883 out.append(sequenceFeature.score);
892 out.append("ENDGROUP").append(TAB);
902 } while (groupIndex < groups.size() + 1);
906 return "No Features Visible";
909 return out.toString();
913 * Parse method that is called when a GFF file is dragged to the desktop
918 AlignViewportI av = getViewport();
921 if (av.getAlignment() != null)
923 dataset = av.getAlignment().getDataset();
927 // working in the applet context ?
928 dataset = av.getAlignment();
933 dataset = new Alignment(new SequenceI[] {});
936 boolean parseResult = parse(dataset, null, false, true);
939 // pass error up somehow
943 // update viewport with the dataset data ?
947 setSeqs(dataset.getSequencesArray());
952 * Implementation of unused abstract method
954 * @return error message
957 public String print()
959 return "Use printGffFormat() or printJalviewFormat()";
963 * Returns features output in GFF2 format, including hidden and non-positional
967 * the sequences whose features are to be output
969 * a map whose keys are the type names of visible features
972 public String printGffFormat(SequenceI[] sequences,
973 Map<String, FeatureColourI> visible)
975 return printGffFormat(sequences, visible, true, true);
979 * Returns features output in GFF2 format
982 * the sequences whose features are to be output
983 * @param featureColours
984 * a map whose keys are the type names of visible features
985 * @param outputVisibleOnly
986 * @param includeNonPositionalFeatures
989 public String printGffFormat(SequenceI[] sequences,
990 Map<String, FeatureColourI> featureColours,
991 boolean outputVisibleOnly,
992 boolean includeNonPositionalFeatures)
994 StringBuilder out = new StringBuilder(256);
995 int version = gffVersion == 0 ? 2 : gffVersion;
996 out.append(String.format("%s %d\n", GFF_VERSION, version));
999 for (SequenceI seq : sequences)
1001 SequenceFeature[] features = seq.getSequenceFeatures();
1002 if (features != null)
1004 for (SequenceFeature sf : features)
1006 isnonpos = sf.begin == 0 && sf.end == 0;
1007 if (!includeNonPositionalFeatures && isnonpos)
1010 * ignore non-positional features if not wanted
1014 // TODO why the test !isnonpos here?
1015 // what about not visible non-positional features?
1016 if (!isnonpos && outputVisibleOnly
1017 && !featureColours.containsKey(sf.type))
1020 * ignore not visible features if not wanted
1025 source = sf.featureGroup;
1028 source = sf.getDescription();
1031 out.append(seq.getName());
1035 out.append(sf.type);
1037 out.append(sf.begin);
1041 out.append(sf.score);
1044 int strand = sf.getStrand();
1045 out.append(strand == 1 ? "+" : (strand == -1 ? "-" : "."));
1048 String phase = sf.getPhase();
1049 out.append(phase == null ? "." : phase);
1051 // miscellaneous key-values (GFF column 9)
1052 String attributes = sf.getAttributes();
1053 if (attributes != null)
1055 out.append(TAB).append(attributes);
1058 out.append(newline);
1063 return out.toString();
1067 * Returns a mapping given list of one or more Align descriptors (exonerate
1070 * @param alignedRegions
1071 * a list of "Align fromStart toStart fromCount"
1072 * @param mapIsFromCdna
1073 * if true, 'from' is dna, else 'from' is protein
1075 * either 1 (forward) or -1 (reverse)
1077 * @throws IOException
1079 protected MapList constructCodonMappingFromAlign(
1080 List<String> alignedRegions, boolean mapIsFromCdna, int strand)
1085 throw new IOException(
1086 "Invalid strand for a codon mapping (cannot be 0)");
1088 int regions = alignedRegions.size();
1089 // arrays to hold [start, end] for each aligned region
1090 int[] fromRanges = new int[regions * 2]; // from dna
1091 int[] toRanges = new int[regions * 2]; // to protein
1092 int fromRangesIndex = 0;
1093 int toRangesIndex = 0;
1095 for (String range : alignedRegions)
1098 * Align mapFromStart mapToStart mapFromCount
1099 * e.g. if mapIsFromCdna
1100 * Align 11270 143 120
1102 * 120 bases from pos 11270 align to pos 143 in peptide
1103 * if !mapIsFromCdna this would instead be
1104 * Align 143 11270 40
1106 String[] tokens = range.split(" ");
1107 if (tokens.length != 3)
1109 throw new IOException("Wrong number of fields for Align");
1116 fromStart = Integer.parseInt(tokens[0]);
1117 toStart = Integer.parseInt(tokens[1]);
1118 fromCount = Integer.parseInt(tokens[2]);
1119 } catch (NumberFormatException nfe)
1121 throw new IOException("Invalid number in Align field: "
1122 + nfe.getMessage());
1126 * Jalview always models from dna to protein, so adjust values if the
1127 * GFF mapping is from protein to dna
1132 int temp = fromStart;
1133 fromStart = toStart;
1136 fromRanges[fromRangesIndex++] = fromStart;
1137 fromRanges[fromRangesIndex++] = fromStart + strand * (fromCount - 1);
1140 * If a codon has an intron gap, there will be contiguous 'toRanges';
1141 * this is handled for us by the MapList constructor.
1142 * (It is not clear that exonerate ever generates this case)
1144 toRanges[toRangesIndex++] = toStart;
1145 toRanges[toRangesIndex++] = toStart + (fromCount - 1) / 3;
1148 return new MapList(fromRanges, toRanges, 3, 1);
1152 * Parse a GFF format feature. This may include creating a 'dummy' sequence to
1153 * hold the feature, or for its mapped sequence, or both, to be resolved
1154 * either later in the GFF file (##FASTA section), or when the user loads
1155 * additional sequences.
1159 * @param relaxedIdMatching
1163 protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment,
1164 boolean relaxedIdMatching, List<SequenceI> newseqs)
1167 * GFF: seqid source type start end score strand phase [attributes]
1169 if (gffColumns.length < 5)
1171 System.err.println("Ignoring GFF feature line with too few columns ("
1172 + gffColumns.length + ")");
1177 * locate referenced sequence in alignment _or_
1178 * as a forward or external reference (SequenceDummy)
1180 String seqId = gffColumns[0];
1181 SequenceI seq = findSequence(seqId, alignment, newseqs,
1184 SequenceFeature sf = null;
1185 GffHelperI helper = GffHelperFactory.getHelper(gffColumns);
1190 sf = helper.processGff(seq, gffColumns, alignment, newseqs,
1194 seq.addSequenceFeature(sf);
1195 while ((seq = alignment.findName(seq, seqId, true)) != null)
1197 seq.addSequenceFeature(new SequenceFeature(sf));
1200 } catch (IOException e)
1202 System.err.println("GFF parsing failed with: " + e.getMessage());
1211 * Process the 'column 9' data of the GFF file. This is less formally defined,
1212 * and its interpretation will vary depending on the tool that has generated
1218 protected void processGffColumnNine(String attributes, SequenceFeature sf)
1220 sf.setAttributes(attributes);
1223 * Parse attributes in column 9 and add them to the sequence feature's
1224 * 'otherData' table; use Note as a best proxy for description
1226 char nameValueSeparator = gffVersion == 3 ? '=' : ' ';
1227 // TODO check we don't break GFF2 values which include commas here
1228 Map<String, List<String>> nameValues = GffHelperBase
1229 .parseNameValuePairs(attributes, ";", nameValueSeparator, ",");
1230 for (Entry<String, List<String>> attr : nameValues.entrySet())
1232 String values = StringUtils.listToDelimitedString(attr.getValue(),
1234 sf.setValue(attr.getKey(), values);
1235 if (NOTE.equals(attr.getKey()))
1237 sf.setDescription(values);
1243 * After encountering ##fasta in a GFF3 file, process the remainder of the
1244 * file as FAST sequence data. Any placeholder sequences created during
1245 * feature parsing are updated with the actual sequences.
1249 * @throws IOException
1251 protected void processAsFasta(AlignmentI align, List<SequenceI> newseqs)
1257 } catch (IOException q)
1260 FastaFile parser = new FastaFile(this);
1261 List<SequenceI> includedseqs = parser.getSeqs();
1263 SequenceIdMatcher smatcher = new SequenceIdMatcher(newseqs);
1266 * iterate over includedseqs, and replacing matching ones with newseqs
1267 * sequences. Generic iterator not used here because we modify
1268 * includedseqs as we go
1270 for (int p = 0, pSize = includedseqs.size(); p < pSize; p++)
1272 // search for any dummy seqs that this sequence can be used to update
1273 SequenceI includedSeq = includedseqs.get(p);
1274 SequenceI dummyseq = smatcher.findIdMatch(includedSeq);
1275 if (dummyseq != null && dummyseq instanceof SequenceDummy)
1277 // probably have the pattern wrong
1278 // idea is that a flyweight proxy for a sequence ID can be created for
1279 // 1. stable reference creation
1280 // 2. addition of annotation
1281 // 3. future replacement by a real sequence
1282 // current pattern is to create SequenceDummy objects - a convenience
1283 // constructor for a Sequence.
1284 // problem is that when promoted to a real sequence, all references
1285 // need to be updated somehow. We avoid that by keeping the same object.
1286 ((SequenceDummy) dummyseq).become(includedSeq);
1287 dummyseq.createDatasetSequence();
1290 * Update mappings so they are now to the dataset sequence
1292 for (AlignedCodonFrame mapping : align.getCodonFrames())
1294 mapping.updateToDataset(dummyseq);
1298 * replace parsed sequence with the realised forward reference
1300 includedseqs.set(p, dummyseq);
1303 * and remove from the newseqs list
1305 newseqs.remove(dummyseq);
1310 * finally add sequences to the dataset
1312 for (SequenceI seq : includedseqs)
1314 // experimental: mapping-based 'alignment' to query sequence
1315 AlignmentUtils.alignSequenceAs(seq, align,
1316 String.valueOf(align.getGapCharacter()), false, true);
1318 // rename sequences if GFF handler requested this
1319 // TODO a more elegant way e.g. gffHelper.postProcess(newseqs) ?
1320 SequenceFeature[] sfs = seq.getSequenceFeatures();
1323 String newName = (String) sfs[0].getValue(GffHelperI.RENAME_TOKEN);
1324 if (newName != null)
1326 seq.setName(newName);
1329 align.addSequence(seq);
1334 * Process a ## directive
1340 * @throws IOException
1342 protected void processGffPragma(String line,
1343 Map<String, String> gffProps, AlignmentI align,
1344 List<SequenceI> newseqs) throws IOException
1347 if ("###".equals(line))
1349 // close off any open 'forward references'
1353 String[] tokens = line.substring(2).split(" ");
1354 String pragma = tokens[0];
1355 String value = tokens.length == 1 ? null : tokens[1];
1357 if ("gff-version".equalsIgnoreCase(pragma))
1363 // value may be e.g. "3.1.2"
1364 gffVersion = Integer.parseInt(value.split("\\.")[0]);
1365 } catch (NumberFormatException e)
1371 else if ("sequence-region".equalsIgnoreCase(pragma))
1373 // could capture <seqid start end> if wanted here
1375 else if ("feature-ontology".equalsIgnoreCase(pragma))
1377 // should resolve against the specified feature ontology URI
1379 else if ("attribute-ontology".equalsIgnoreCase(pragma))
1381 // URI of attribute ontology - not currently used in GFF3
1383 else if ("source-ontology".equalsIgnoreCase(pragma))
1385 // URI of source ontology - not currently used in GFF3
1387 else if ("species-build".equalsIgnoreCase(pragma))
1389 // save URI of specific NCBI taxon version of annotations
1390 gffProps.put("species-build", value);
1392 else if ("fasta".equalsIgnoreCase(pragma))
1394 // process the rest of the file as a fasta file and replace any dummy
1396 processAsFasta(align, newseqs);
1400 System.err.println("Ignoring unknown pragma: " + line);